Title: | Stan Routines For Univariate And Multivariate Time Series |
---|---|
Description: | Bundles univariate and multivariate STAN scripts for FISH 507 class. |
Authors: | Eric J. Ward [aut, cre], Mark D. Scheuerell [aut], Elizabeth E. Holmes [aut], Kiva L. Oken [aut], Trustees of Columbia University [cph] |
Maintainer: | Eric J. Ward <[email protected]> |
License: | GPL (>=3) |
Version: | 0.1.6 |
Built: | 2024-11-08 05:56:15 UTC |
Source: | https://github.com/atsa-es/atsar |
A DESCRIPTION OF THE PACKAGE
Stan Development Team (2020). RStan: the R interface to Stan. R package version 2.21.2. https://mc-stan.org
fit_stan is the primary function which calls pre-written stan scripts for time series data.
fit_stan( y, x = NA, model_name = NA, est_drift = FALSE, est_mean = FALSE, P = 1, Q = 1, mcmc_list = list(n_mcmc = 1000, n_burn = 500, n_chain = 3, n_thin = 1), family = "gaussian", est_nu = FALSE, marss = list(states = NULL, obsVariances = NULL, proVariances = NULL, trends = NULL), map_estimation = FALSE, hessian = FALSE, ... )
fit_stan( y, x = NA, model_name = NA, est_drift = FALSE, est_mean = FALSE, P = 1, Q = 1, mcmc_list = list(n_mcmc = 1000, n_burn = 500, n_chain = 3, n_thin = 1), family = "gaussian", est_nu = FALSE, marss = list(states = NULL, obsVariances = NULL, proVariances = NULL, trends = NULL), map_estimation = FALSE, hessian = FALSE, ... )
y |
The response variable (numeric) |
x |
The predictors, either a vector or matrix |
model_name |
The specific name of the model to be fitted. Currently supported are 'regression', 'ar', 'rw', 'ma', 'ss_ar' (state space univariate AR), or 'ss_rw' (state space univariate random walk). |
est_drift |
Whether or not to estimate a drift parameter (default = FALSE). Only applicable to the rw and ar models. |
est_mean |
Whether to estimate a mean or not (for state space autoregressive model only) |
P |
The order of the ar model, with minimum value = 1 (default). |
Q |
The order of the ma model, with minimum value = 1 (default). |
mcmc_list |
A list of MCMC control parameters. These include the number of 'iterations' (default = 1000), burn in or warmup (default = 500), chains (default = 3), and thinning (default = 1) |
family |
A named distribution for the observation model, defaults to gaussian |
est_nu |
Boolean, whether to model process deviations as Student-t or not (default). |
marss |
A named list containing the following elements for specifying marss models: (states=NULL, obsVariances=NULL, proVariances=NULL, trends=NULL |
map_estimation |
Whether to do maximum a posteriori estimation via [rstan::optimizing()] (defualts to FALSE) |
hessian |
Whether to return hessian if map_estimation is TRUE via [rstan::optimizing()] |
... |
Any other arguments passed to [rstan::sampling()]. |
an object of class 'rstan'